Fuzzy Support Vector Machine Method for Evaluating Innovation Sources in Service Firms

Abstract

The traditional evaluating methods can not deal with the evaluation problem of innovation sources in service firms with fuzzy information, authors have constructed fuzzy support vector machine based on support vector machine and fuzzy chance constrained programming, and applied this new method to evaluating innovation sources in service firms. In basis of related literature reviewing, authors have summarized 7 indicators of innovation sources in service firms which are internal R&D, staff quality, customers, suppliers and so on. By selecting the data of 80 service firms of Zhejiang Province in China as samples and 60 firms being as training samples, we obtain the model of evaluating innovation sources of service firms. Simultaneously the other 20 firms being as testing samples, the results show higher accuracy rate (90%) and lower error rate(average error is 0.038). Therefore fuzzy support vector machine method for innovation sources in service firms provides a new way in the process of selecting innovation sources.